Home
last modified time | relevance | path

Searched +full:nvidia +full:- +full:smi (Results 1 – 25 of 33) sorted by relevance

12

/external/pytorch/.github/actions/linux-test/
Daction.yml1 name: linux-test
4 build-environment:
7 description: Top-level label for what's being built/tested.
8 test-matrix:
12 docker-image:
16 sync-tag:
22 job with the same `sync-tag` is identical.
23 use-gha:
28 dashboard-tag:
32 s3-bucket:
[all …]
/external/pytorch/.github/workflows/
D_linux-test.yml1 name: linux-test
6 build-environment:
9 description: Top-level label for what's being built/tested.
10 test-matrix:
14 docker-image:
18 sync-tag:
24 job with the same `sync-tag` is identical.
25 timeout-minutes:
31 use-gha:
36 dashboard-tag:
[all …]
/external/pytorch/.ci/pytorch/
Dwin-test.sh2 set -ex
9 TMP_DIR_WIN=$(cygpath -w "${TMP_DIR}")
12 PROJECT_DIR_WIN=$(cygpath -w "${PROJECT_DIR}")
15 TEST_DIR_WIN=$(cygpath -w "${TEST_DIR}")
17 export PYTORCH_FINAL_PACKAGE_DIR="${PYTORCH_FINAL_PACKAGE_DIR:-/c/w/build-results}"
18 PYTORCH_FINAL_PACKAGE_DIR_WIN=$(cygpath -w "${PYTORCH_FINAL_PACKAGE_DIR}")
21 mkdir -p "$TMP_DIR"/build/torch
23 export SCRIPT_HELPERS_DIR=$SCRIPT_PARENT_DIR/win-test-helpers
38 python -m pip install pytest-rerunfailures==10.3 pytest-cpp==2.3.0 tensorboard==2.13.0
41 python -m pip install z3-solver==4.12.2.0
[all …]
/external/tensorflow/tensorflow/python/distribute/cluster_resolver/
Dslurm_cluster_resolver.py1 # Copyright 2018-2020 The TensorFlow Authors. All Rights Reserved.
7 # http://www.apache.org/licenses/LICENSE-2.0
31 Input: 'n[1-2],m5,o[3-4,6,7-9]')
36 """Split hostlist at commas outside of range expressions ('[3-5]')."""
56 """Expand a range expression like '3-5' to values 3,4,5."""
58 sub_range = part.split('-')
101 raise ValueError('Invalid tasks-per-node list format "%s": %s' %
135 """Gets the number of NVIDIA GPUs by using CUDA_VISIBLE_DEVICES and nvidia-smi.
140 RuntimeError if executing nvidia-smi failed
145 pass # Ignore and fallback to using nvidia-smi
[all …]
DREADME_Slurm.md36 fallback to using `nvidia-smi`. If this doesn't work or non-NVIDIA GPUs are used
43 - Slurm allocation in shell `salloc --nodes=2 -t 01:30:00 --ntasks-per-node=2
44 --gres=gpu:k80:4 --exclusive`
45 - Run the example `srun python tf_example.py`
46 - Creating cluster in Python `import tensorflow as tf cluster_resolver =
64 - Assuming the same job parameters (`salloc` & `srun`) as above
65 - Creating cluster in Python ``` cluster_resolver =
90 - `_resolve_own_rank`
91 - `_resolve_num_tasks`
92 - `_resolve_hostlist`
[all …]
/external/pytorch/torch/utils/
Dcollect_env.py1 # mypy: allow-untyped-defs
5 # Run it with `python collect_env.py` or `python -m torch.utils.collect_env`
73 """Return (return-code, stdout, stderr)."""
130 return run_and_parse_first_match(run_lambda, 'gcc --version', r'gcc (.*)')
133 return run_and_parse_first_match(run_lambda, 'clang --version', r'clang version (.*)')
137 return run_and_parse_first_match(run_lambda, 'cmake --version', r'cmake (.*)')
142 cmd = 'kextstat | grep -i cuda'
144 r'com[.]nvidia[.]CUDA [(](.*?)[)]')
145 smi = get_nvidia_smi()
146 return run_and_parse_first_match(run_lambda, smi, r'Driver Version: (.*?) ')
[all …]
/external/pytorch/.devcontainer/
DREADME.md9 3. Run the installer and follow the on-screen instructions to install VSCode on your system.
18 …[homepage](https://marketplace.visualstudio.com/items?itemName=ms-vscode-remote.remote-containers)…
22 …ocs.docker.com/get-docker/) to install Docker. Don't forget the [post installation steps](https://…
24 If you are using [Visual Studio Code Remote - SSH](https://code.visualstudio.com/docs/remote/ssh), …
26 ## Step 4 (Optional): Install NVIDIA Container Toolkit for GPU Usage
28 …d to use GPU resources, first ensure you have NVIDIA drivers installed on your system. Check if `n…
29 …l guide](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.ht…
32 docker run --rm --runtime=nvidia --gpus all nvidia/cuda:11.6.2-base-ubuntu20.04 nvidia-smi
72 For an in-depth understanding of Dev Container and its caveats, please refer to [the full documenta…
/external/coreboot/src/mainboard/acer/aspire_vn7_572g/acpi/
Dec.asl1 /* SPDX-License-Identifier: GPL-2.0-only */
5 * - TRPS: This is SMI 0xDD, likely in SmmOemDriver. This SW SMI adds to and executes
6 * a table of function pointers produced throughout the OEM 'value-add' stack.
7 * - Arg0 - "SFUN" - is index into "$FNC" pointer table? It's easier to
9 * - Known functions:
10 * - 0x80 calls offset 0 in ACER_BOOT_DEVICE_SERVICE_PROTOCOL_GUID.
11 * - NB: efiXplorer can miss InstallProtocolInterface() when Interface is local
12 * - 0x81 toggles Intel Dynamic Acceleration in IA32_MISC_ENABLE MSR.
13 * - 0x82 does switch on "OSYS" to set EC byte. Suspect this is for OS features.
15 * - RBEC/WBEC/MBEC: This is SMI 0xDD, "functions" 0x10, 0x11 and 0x12 in SmmKbcDriver,
[all …]
/external/angle/src/tests/
Drun_perf_tests.py4 # Use of this source code is governed by a BSD-style license that can be
79 pattern = r'\.' + result + r':.*= ([0-9.]+)'
93 return sorted(data)[n:-n]
106 ss = sum((float(x) - c)**2 for x in data)
255 '--trials',
260 run_args += ['--steps-per-trial', str(steps_per_trial)]
262 run_args += ['--trial-time', str(args.trial_time)]
265 run_args += ['--warmup'] # Render each frame once with glFinish
268 run_args += ['--perf-counters', args.perf_counters]
272 run_args += ['--isolated-script-test-perf-output=%s' % histogram_file_path]
[all …]
/external/pytorch/benchmarks/distributed/ddp/
Dbenchmark.py6 # b) an increasing number of processes. This produces a 1-GPU baseline,
7 # an 8-GPU baseline (if applicable), as well as measurements for however
38 return allgather_object(proc.stdout.decode("utf-8"))
57 optimizer = optim.SGD(model.parameters(), 0.001, momentum=0.9, weight_decay=1e-4)
77 measurements.append(time.time() - start)
103 prefix = f"{len(ranks):4} GPUs -- {prefix}"
134 # Multi-machine benchmarks
196 parser.add_argument("--rank", type=int, default=os.environ["RANK"])
197 parser.add_argument("--world-size", type=int, required=True)
198 parser.add_argument("--distributed-backend", type=str, default="nccl")
[all …]
/external/tensorflow/tools/
Dtf_env_collect.sh8 # http://www.apache.org/licenses/LICENSE-2.0
17 set -u # Check for undefined variables
93 num=`cat /proc/1/cgroup | grep docker | wc -l`;
94 if [ $num -ge 1 ]; then
102 c++ --version 2>&1
111 ${python_bin_path} -c "import sys;print(hasattr(sys, \"real_prefix\"))"
127 LD_DEBUG=libs ${python_bin_path} -c "import tensorflow" 2>>${OUTPUT_FILE} > /tmp/loadedlibs
133 if [ -z ${LD_LIBRARY_PATH+x} ]; then
138 if [ -z ${DYLD_LIBRARY_PATH+x} ]; then
146 echo '== nvidia-smi ==================================================='
[all …]
/external/tensorflow/tensorflow/tools/ci_build/linux/ppc64le/gpu/
Drun_py2.sh8 # http://www.apache.org/licenses/LICENSE-2.0
18 set -e
19 set -x
21 N_JOBS=$(grep -c ^processor /proc/cpuinfo)
22 LT_JOBS=$(nvidia-smi --query-gpu=gpu_name --format=csv,noheader | wc -l)
31 export CC_OPT_FLAGS='-mcpu=power8 -mtune=power8'
39 bazel test --config=cuda --test_tag_filters=-no_oss,-oss_serial,-no_gpu,-benchmark-test -k \
40 --jobs=${N_JOBS} --test_timeout 300,450,1200,3600 \
41 --test_output=errors --local_test_jobs=${LT_JOBS} --build_tests_only --config=opt \
42 --test_size_filters=small,medium \
[all …]
Drun_py3.sh8 # http://www.apache.org/licenses/LICENSE-2.0
18 set -e
19 set -x
21 N_JOBS=$(grep -c ^processor /proc/cpuinfo)
22 LT_JOBS=$(nvidia-smi --query-gpu=gpu_name --format=csv,noheader | wc -l)
31 export CC_OPT_FLAGS='-mcpu=power8 -mtune=power8'
39 bazel test --config=cuda --test_tag_filters=-no_oss,-oss_serial,-no_gpu,-benchmark-test -k \
40 --jobs=${N_JOBS} --test_timeout 300,450,1200,3600 \
41 --test_output=errors --local_test_jobs=${LT_JOBS} --build_tests_only --config=opt \
42 --test_size_filters=small,medium \
[all …]
/external/pytorch/benchmarks/inference/
DREADME.md6 ResNet-18 checkpoint to 'cuda:0' and compiles the model. It accepts requests in
17 3. A thread that polls nvidia-smi for GPU utilization metrics.
19 For now we omit data preprocessing as well as result post-processing.
24 - `num_iters` (default: 100): how many requests to send to the backend
26 - `batch_size` (default: 32): the batch size of the requests.
27 - `model_dir` (default: '.'): the directory to load the checkpoint from
28 - `compile` (default: compile): or `--no-compile` whether to `torch.compile()`
30 …- `output_file` (default: output.csv): The name of the csv file to write the outputs to in the `re…
31 …- `num_workers` (default: 2): The `max_threads` passed to the `ThreadPoolExecutor` in charge of mo…
36 python -W ignore server.py --num_iters 1000 --batch_size 32
Dserver.py57 warmup_response_time = time.time() - request_time
59 response_times.append(time.time() - request_time)
71 self.end_recv_time - self.start_send_time
76 This function will poll nvidia-smi for GPU utilization every 100ms to
84 "nvidia-smi",
85 "--query-gpu=utilization.gpu",
86 "--id=0",
87 "--format=csv,noheader,nounits",
200 f"{self.model_dir}/resnet18-f37072fd.pth",
204 self.metrics_dict["torch_load_time"] = time.time() - start_load_time
[all …]
/external/pytorch/torch/distributed/benchmarks/
Dbenchmark_ddp_rpc.py1 # mypy: allow-untyped-defs
3 # pyre-unsafe
76 emb_lookups_reshaped = emb_lookups_cat.reshape( # type: ignore[possibly-undefined]
109 …proc = subprocess.run(shlex.split(cmd), capture_output=True, check=False) # type: ignore[call-ove…
113 torch.save(proc.stdout.decode("utf-8"), buffer)
175 # Include warm-up cycles during training
197 measurements.append(time.time() - start)
200 # Throw away warm-up measurements
202 return rank, measurements, batch_size # type: ignore[possibly-undefined]
219 backend=BackendType.TENSORPIPE, # type: ignore[attr-defined]
[all …]
DREADME.md20 …sing [Distributed Autograd](https://pytorch.org/docs/main/rpc.html#distributed-autograd-framework).
25 7) Finally, the [Distributed Optimizer](https://pytorch.org/docs/main/rpc.html#module-torch.distrib…
30 ---------- Info ---------
35 ---------- nvidia-smi topo -m ---------
38 GPU0 X NV2 NV1 NV2 NV1 NODE NODE NODE 0-19,40-59
39 GPU1 NV2 X NV2 NV1 NODE NV1 NODE NODE 0-19,40-59
40 GPU2 NV1 NV2 X NV1 NODE NODE NV2 NODE 0-19,40-59
41 GPU3 NV2 NV1 NV1 X NODE NODE NODE NV2 0-19,40-59
42 GPU4 NV1 NODE NODE NODE X NV2 NV1 NV2 0-19,40-59
43 GPU5 NODE NV1 NODE NODE NV2 X NV2 NV1 0-19,40-59
[all …]
/external/coreboot/src/mainboard/lenovo/t520/
Ddevicetree.cb30 device ref peg10 on end # NVIDIA Copcie_rporation GF119M [NVS 4200M]
38 # 1 SMI# (if corresponding ALT_GPI_SMI_EN bit is also set)
53 # Enable zero-based linear PCIe root port functions
/external/pytorch/torch/cuda/
D__init__.py1 # mypy: allow-untyped-defs
11 :ref:`cuda-semantics` has more details about working with CUDA.
41 from torch._C import _cudart # type: ignore[attr-defined]
85 def _exchange_device(device: int) -> int:
87 return -1
95 def _maybe_exchange_device(device: int) -> int:
97 return -1
107 def _is_compiled() -> bool:
112 def _nvml_based_avail() -> bool:
116 def is_available() -> bool:
[all …]
/external/coreboot/src/mainboard/lenovo/t420s/
Ddevicetree.cb31 device ref peg10 on end # NVIDIA Copcie_rporation GF119M [NVS 4200M]
39 # 1 SMI# (if corresponding ALT_GPI_SMI_EN bit is also set)
56 # Enable zero-based linear PCIe root port functions
64 {0, 1, -1}, /* P0: empty */
66 {1, 1, -1}, /* P2: HALF MINICARD (WLAN) no oc */
67 {1, 0, -1}, /* P3: WWAN, no OC */
68 {1, 1, -1}, /* P4: smartcard, no OC */
69 {1, 1, -1}, /* P5: ExpressCard, no OC */
70 {0, 0, -1}, /* P6: empty */
71 {0, 0, -1}, /* P7: empty */
[all …]
/external/tensorflow/tensorflow/core/util/
Dtest_log.proto117 string uuid = 2; // Final entry in output of "nvidia-smi -L"
125 string release = 4; // e.g. '3.13.0-76-generic'
127 string version = 6; // e.g. '#120-Ubuntu SMP Mon Jan 18 15:59:10 UTC 2016'
159 // Run-specific items such as arguments to the test / benchmark.
193 // Machine-specific parameters (Platform and CPU info)
196 // Run-specific parameters (arguments, etc)
/external/pytorch/docs/source/notes/
Dhip.rst1 .. _hip-semantics:
6 ROCm\ |trade| is AMD’s open source software platform for GPU-accelerated high
10 projects that require portability between AMD and NVIDIA.
15 ----------------------------------------
21 The example from :ref:`cuda-semantics` will work exactly the same for HIP::
25 cuda2 = torch.device('cuda:2') # GPU 2 (these are 0-indexed)
60 ----------------
77 TensorFloat-32(TF32) on ROCm
78 ----------------------------
82 .. _rocm-memory-management:
[all …]
Dfaq.rst5 -------------------------------------------------------
20 Sometimes, it can be non-obvious when differentiable variables can
22 <https://discuss.pytorch.org/t/high-memory-usage-while-training/162>`_):
24 .. code-block:: python
40 `1 <https://discuss.pytorch.org/t/resolved-gpu-out-of-memory-error-with-batch-size-1/3719>`_.
53 .. code-block:: python
75 `this forum post <https://discuss.pytorch.org/t/help-clarifying-repackage-hidden-in-word-language-m…
86 You can trade-off memory for compute by using `checkpoint <https://pytorch.org/docs/stable/checkpoi…
89 ----------------------------------
91 result, the values shown in ``nvidia-smi`` usually don't reflect the true
[all …]
/external/coreboot/Documentation/releases/
Dcoreboot-4.9-relnotes.md22 --------
43 -------------
50 read-only) in parallel to the [new documentation
55 --------------
66 ------------
84 ---------------
105 -------------------
106 * ASROCK G41C-GS
107 * ASROCK G41M-GS
108 * ASROCK G41M-S3
[all …]
/external/coreboot/util/lint/
Dcoreboot.dict25 adl-m
26 adl-n
27 adl-p
102 auto-demotion
123 base-address
133 big-endian
205 c-state
206 c-states
232 cd-rom
268 clause-patent
[all …]

12